Generating interrelated path groups by using machine learning
US-10846453-B1 · Nov 24, 2020 · US
US2020209832A1 · US · A1
| Field | Value |
|---|---|
| Publication number | US-2020209832-A1 |
| Application number | US-201816236949-A |
| Country | US |
| Kind code | A1 |
| Filing date | Dec 31, 2018 |
| Priority date | Dec 31, 2018 |
| Publication date | Jul 2, 2020 |
| Grant date | — |
A practical reading order for non-experts. Skip the full description unless you need deep technical detail.
What the patent document calls the invention.
A short plain-language summary of the technical disclosure.
Who owns or filed the patent and who is credited as inventor.
Filing, priority, publication, and grant dates set the timeline.
The legal scope of protection — read this for what is actually claimed.
Technology tags used to group this patent with similar filings.
Prior art links and similar publications in this corpus.
Official abstract text for this publication.
The present disclosure is directed to a method and system for hierarchical multi-scale design with the aid of a digital computer. A hierarchical representation of a shape and material distribution is constructed which satisfies a top-level constraint at a top-level of representation. Properties for families of designs at each of the lower levels of representation that satisfy additional constraints link each of the lower levels of representation to at least a next higher level of the representation.
Opening claim text (preview).
1 . A computer-implemented method comprising: specifying at least one design criterion comprising at least one constraint; constructing a hierarchical representation of families of designs that satisfy the specified criteria at every level of the representation, the hierarchical representation comprising one or more levels of representation wherein lower levels of representation provide more details about the designs than the higher levels of representation, thereby further restrict the design families; and synthesizing families of designs at every level of the representation that satisfy the specified criteria as well as additional constraints that link the levels of the representation. 2 . The method of claim 1 , wherein the design criteria further comprise at least one objective function that provides a figure of merit for designs that satisfy the constraints, the synthesizing the families of designs further comprising optimization with respect to the objective function. 3 . The method of claim 1 , wherein the hierarchical representation comprises of surrogate properties of shape and material distributions for the families of designs, the surrogate properties providing sufficient information to evaluate the families of designs against the design criteria. 4 . The method of claim 3 , wherein the surrogate properties comprise finite subsequences of a convergent sequence or a series of variables that represent the projection of the shape and material distributions at a given level onto a functional basis. 5 . The method of claim 4 , wherein a longer subsequence of variables is decided at lower levels of the hierarchical representation. 6 . The method of claim 3 , wherein each level of the hierarchical representation comprises a spatial cellular decomposition wherein every cell at every level of the spatial cellular decomposition is assigned with one or more surrogate properties. 7 . The method of claim 6 , wherein the cells at each of the levels are decomposed into disjoint cells to obtain the cells at a next lower level. 8 . The method of claim 6 , wherein each level of the hierarchical representation describes the shape and material distributions at a size scale determined by the size of the cells. 9 . The method of claim 6 , wherein additional constraints link the properties of a cell at a given level to properties of the cells obtained from the decomposition. 10 . The method of claim 9 , wherein the surrogate properties are integral properties of the shape and material distributions within a cell and the additional constraints that link them across the levels are of additivity of integral properties. 11 . The method of claim 3 , wherein the final shape and material distributions are not determined until the surrogate properties are synthesized for all levels of representation. 12 . A system comprising: a processor coupled to memory, the memory storing instructions executable by the processor for performing: synthesizing a hierarchical representation of surrogate properties of shape and material distributions of a family of designs by a spatial cellular decomposition with properties assigned to each cell at each level of the hierarchical representation; starting from a top level of the hierarchical representation, determining properties at the top level that satisfy constraints specified at the top level based on an analysis performed at the top level; for each level below the top level, performing a process comprising further decomposing each cell at a given level into disjoint cells of the next level, decisions on the surrogate properties assigned to the child cells being made based on an analysis performed at the given level subject to additional constraints that enforce the properties at the cell at the given level to remain consistent with the surrogate properties decided at the cell at the previous level. 13 . The system of claim 12 , wherein the surrogate properties synthesized at each level declaratively represent a family or equivalence class of designs that possess those surrogate properties, and wherein the decisions at each level further restrict the class of designs into subclasses by redistributing the properties of a cell in the previous level. 14 . The system of claim 12 , wherein the process terminates when the family or the equivalence class of designs is restricted enough at the bottom level of the representation to be interpreted as a class of interchangeable designs realized by at least one of given manufacturing processes and material microstructures. 15 . The system of claim 14 , wherein the given manufacturing process or material microstructure is finalized after the synthesis of surrogate properties is complete at all levels. 16 . The system of claim 12 , wherein the processor further optimizes of the surrogate properties at one or more of the levels according to one or more objective functions. 17 . The system of claim 16 , wherein the optimization comprises enumerating solutions along Pareto fronts. 18 . The system of claim 12 , wherein the decisions at each level are performed by iterations over one or more candidate families of designs represented by surrogate properties at the given level. 19 . The system of claim 18 , wherein the iterations are guided by sensitivities or gradients of at least one of the constraints or objective functions at the given level to changes in the surrogate properties. 20 . A method comprising visualizing a hierarchical representation of surrogate properties of shape and material distributions of a family of designs by a spatial cellular decomposition with properties assigned to each cell at each level of the hierarchical representation; and displaying a color-coding of the values of surrogate properties at specified levels of the representation.
Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA] · CPC title
using finite element methods [FEM] or finite difference methods [FDM] · CPC title
Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30) · CPC title
Making, forming 3-D object, model, surface · CPC title
Surface or curve machining, making three-dimensional [3D] objects, e.g. desktop manufacturing · CPC title
Related publications grouped by family.
Answers are generated from the same data shown on this page.